Understand markdown.
Work better with AI.

CLAUDE.md, .cursorrules, subfolder rules — here's what goes in them and how to scope rules to where they matter.

Cursor and Claude Code aren't magic — they're a system prompt and some markdown files. You can build the same thing for your own work, right now.

Markdown is the format AI tools read best. The syntax that matters — and why each element makes your AI interactions sharper.

Every major AI chat interface is built on markdown — and every one of them systematically works against it.

The copy button and Ctrl+C don't give you the same thing. Here's why — and what to do about it.

You don't need a vector database to give an AI agent meaningful background knowledge. A plain Markdown file is faster, more readable, and easier to maintain.

A context file is written once and read once. A knowledge base grows for years. Here's how to structure Markdown notes so they stay findable as the collection grows.

Wikilinks, backlinks, callouts, tags — what Obsidian actually adds on top of plain Markdown, and what an AI agent sees when it reads the files directly.

Jupyter's .ipynb format embeds markdown natively. What that means for AI-generated notebooks and why it matters.

Write the spec first, in Markdown. Then hand it to AI. Cleaner output, fewer correction cycles, and a document that stays useful long after the first session.

Phases, slices, and what happens to documentation after launch — plus three ready-to-use prompts for spec-driven projects.

When you ask an AI to draw a diagram, it almost always writes Mermaid syntax. Here's what Mermaid is, why AI converged on it, and how to render what you get.

Fenced code blocks, language tags, and why the same markdown renders so differently depending on where you put it.

AI writes math with LaTeX syntax. Here's what that means, where it renders, and how to get clean output in your documents.

Plain text with meaning — what the symbols do, why it became the default format of the web, and why AI speaks it natively.

Why structured plain text is the most reliable format for both writing prompts and receiving AI output.

How a simple tool for converting plain text to HTML became the default language of the internet — and then of AI.

The distinction between a document title and its internal headings — and why it matters more when AI is reading it.

CommonMark, GFM, Obsidian — for AI tools there is one clear choice, and a clear reason why.